₹ 70000
Budget
19
Proposals
1430
Views
Active
Status
Skills Required
Project Details
Need an Expert to start the project with first identifying a light weight LLM fit for my Use Case, which is to generate text summarizations and Q&A on proprietary Book on Finance (containing text, tables, images, basic calculations). The foundation model (may be DistilBERT) need to be further Quantized using QLoRA - to reduce weight matrix and reduction in compute cost.
This Quantized model shall be pruned, fine-tuned & trained on our proprietary Book data using techniques of prompt Engineering, RAG, LangChain, FAISS etc. - such that the ChatBot is able to generate accurate & relevant Responses & Summaries at least compute cost. We will test this model using StreamLit and Colab/Kaggle, and monitor the latency, number count, compute cost etc. on a Dashboard. Once testing results are satisfactory we would deploy it on Cloud, with a web endpoint for Use.
Attachments
1699671072-Custom-ChatBot-Development.pdf
About the Client
India
Member since
Copyright © 2025 | Truelancer.com